Virology 433 (2012) 64–72

Contents lists available at SciVerse ScienceDirect

Virology

0042-68

http://d

n Corr

Microbi

TX 7783

E-m

journal homepage: www.elsevier.com/locate/yviro

Human cytomegalovirus infection of THP-1 derived macrophages revealsstrain-specific regulation of actin dynamics

V. Sanchez n, J.J. Dong, J. Battley, K.N. Jackson, B.C. Dykes

Texas A&M Health Science Center, College of Medicine, Department of Microbial and Molecular Pathogenesis, College Station, TX 77843, USA

a r t i c l e i n f o

Article history:

Received 26 January 2012

Returned to author for revisions

21 May 2012

Accepted 16 July 2012Available online 5 August 2012

Keywords:

Human cytomegalovirus

Macrophage differentiation

Replication

Protein trafficking

Actin cytoskeleton

22/$ - see front matter & 2012 Elsevier Inc. A

x.doi.org/10.1016/j.virol.2012.07.015

espondence to: Texas A&M Health Science C

al and Molecular Pathogenesis, SRPH Bldg A,

3-1266, USA. Fax: þ1 979 862 7226.

ail address: VSanchez@medicine.tamhsc.edu

a b s t r a c t

Human cytomegalovirus (HCMV) remains latent in cells of the myeloid lineage after primary infection.

The THP-1 monocytic cell line is conditionally permissive for infection and has been used primarily to

study the process of HCMV reactivation when the cells are induced to differentiate. In the present

report, we characterized lytic infection in THP-1 derived macrophages using two strains of HCMV,

Towne and BAC-derived TR. Our findings indicate that these cells express viral genes of all three kinetic

classes and produce extracellular virus, but that there is a delay in these processes relative to

productively infected fibroblasts. Importantly, our studies in THP-1 derived macrophages revealed

strain-specific differences in pp65 trafficking and actin dynamics. Based on these observations, our

studies indicate that differentiated THP-1 cells can serve as a valuable model for lytic infection.

& 2012 Elsevier Inc. All rights reserved.

Introduction

Human cytomegalovirus is a large beta-herpesvirus that is acommon, opportunistic pathogen. Infections are generally asymp-tomatic but the virus remains latent in the host after primaryinfection. While the virus replicates in numerous cell types, thereservoir for the virus in the body appears to be cells of themyeloid lineage. Work by Soderberg-Naucler and coworkers firstidentified peripheral blood mononuclear cells as the reservoir oflatent virus (Soderberg-Naucler et al., 1997). A number of morerecent studies have aimed at identifying the earliest cellulardevelopmental stage that contains latent viral genomes. HCMVhas been shown to establish latency in bone marrow CD34þmyeloid progenitor cells (for review, see (Reeves and Sinclair,2008)). Upon terminal differentiation into macrophages or den-dritic cells, the latent virus is reactivated and replicates.

Cells of mononuclear phagocyte system are important in bothinnate and acquired immune responses to infection (for review,see (Martinez et al., 2006; Murray and Wynn, 2011)). The systemis comprised of myeloid progenitor cells in the bone marrow,circulating monocytes, tissue macrophages and dendritic cellsdistributed throughout the body. In addition to providing the firstline of defense in the clearance of microbes by phagocytosis,macrophages and dendritic cells present peptide antigens that

ll rights reserved.

enter, Department of

Rm. 234, College Station,

(V. Sanchez).

prime the acquired immune response. Monocyte differentiationinto macrophages is driven by GM-CSF, M-CSF, and CSF-1. Thefactors activate pro-survival pathways and induce changes ingene expression driving the morphological changes associatedwith the macrophage phenotype. Primary monocytes isolatedfrom peripheral blood can be used to investigate the process ofdifferentiation in culture; however, because these cells are short-lived, in vitro studies necessitate routine isolation of the cells. Thestudy of differentiation has been greatly facilitated by the avail-ability of human myelomonocytic cell lines such as U937, HL60,and THP-1 that can be induced to differentiate into macrophage-like cells with GM-CSF, M-CSF, retinoic acid, or phorbol esters.

We have chosen to work with THP-1 monocytic cells that lacka functional p53 (Akashi et al., 1999). Numerous studies haveutilized THP-1 cells to investigate the process of monocytedifferentiation setting the precedent for our use of this cell line(Auwerx, 1991; Kohro et al., 2004). Upon treatment with phorbolesters, THP-1 monocytes develop macrophage-like features bothmorphologically and functionally. THP-1 cells have previouslybeen described as conditionally permissive for HCMV infectionand IE gene expression can be induced after differentiation of thecells. Most studies have used the cells to study viral reactivation(Ioudinkova et al., 2006; Keyes et al., 2012; Saffert and Kalejta,2010; Sinclair et al., 1992; Yeeet al., 2007); however, Turtinen andSeufzer investigated lytic infection in these cells and showed thatthe THP-1 derived macrophages supported replication of thelaboratory strain Towne, but not AD169 (Turtinen and Seufzer,1994). A recent report by McCormick and colleagues has extendedthe characterization of THP-1 derived macrophages during HCMV

Table 1Percentage HCMV-infected THP-1 derived macrophages at

48 h p.i.

Expt. Virus %IEþ %UL44þ %UL44þ /IEþ

1 Towne 70 63 90

TR-BAC 72 71 99

2 Towne 77 67 87

TR-BAC 86 75 87

3 Towne 74 71 96

TR-BA 85 68 81

4 Towne 86 73 85

TR-BAC 86 70 81

5 Towne 75 60 79

TR-BAC 68 57 84

V. Sanchez et al. / Virology 433 (2012) 64–72 65

lytic infection (McCormick et al., 2010). In their studies, theauthors used THP-1 cells to investigate the role of UL36 incontrolling apoptotic pathways during infection. While theseauthors focused on the effects of infection on cellular pathways,we directed our efforts toward characterizing stages of the virallytic cycle in THP-1 derived macrophages. Our data indicate thatthere is a delay in both peak extracellular virus production andvirus-induced cell death in THP-1 macrophages relative to thekinetics observed during infection of human fibroblasts at a highmultiplicity. Viral proteins of all three kinetic classes areexpressed within 4 days post-infection in the macrophage-likecells but viral DNA replication is slow to increase within this timeperiod. Comparison of two HCMV strains, Towne and BAC-derivedTR (TR-BAC), did not reveal substantial differences in titers, viralDNA replication, or viral gene expression. Interestingly, weobserved a significant delay in the cytoplasmic accumulation ofpp65 in THP-1 derived macrophages. This delay was especiallypronounced in cells infected with the BAC-derived TR strain ofHCMV and this phenotype was later also detected in fibroblastsinfected with this strain. In addition to the differences observed inpp65 localization, we noted that the morphology of THP-1macrophages infected with the two strains of HCMV was strik-ingly different. This observation led us to investigate the distribu-tion of actin filaments in cells infected with either strain. Ourfindings indicate that Towne and TR-BAC infections had distincteffects on the actin cytoskeleton, such that TR-BAC stronglyinduced assembly of actin stress fibers. These filaments were alsoevident in Towne-infected THP-1 macrophages, but at a lowerfrequency. In contrast, Towne-infection of primary fibroblasts ledto depolymerization of actin while stress fibers were maintainedduring TR-BAC infection. Despite these cell type-specific differ-ences in Towne-induced remodeling of the actin cytoskeleton, thestudies presented here validate the use of THP-1 derived macro-phages as a model for lytic infection.

Fig. 1. Extracellular virus production by THP-1 derived macrophages. (A.) THP-1

cells were treated with PMA to induce differentiation for 22–24 h prior to

infection with HCMV Towne (T) or BAC-derived TR (TRB) at a multiplicity of

5 as described in the Materials and Methods. At the time points indicated,

supernatants were collected and frozen. Viral titers were determined by plaque

assay using primary human fibroblasts. Bars represent the mean of values from

three independent experiments indicated by the scatter plot. (B.) Differentiated

THP-1 cells were infected as described in the Materials and Methods. At d7 p.i.,

supernatants and cells were collected. Cells were lysed to release pools of

intracellular virus that were measured by plaque assay as above. Bars represent

the mean of values from two independent experiments indicated by the

scatter plot.

Results

HCMV efficiently infects THP-1 derived macrophages but

extracellular virus production is delayed

For our initial characterization of HCMV infection of differen-tiated THP-1 cells, the efficiency of infection and production ofextracellular virus were determined. THP-1 cells were treated withPMA one day prior to infection to induce differentiation. Twenty-two to 24 h later, the adherent cells were infected at an MOI of5 with either Towne or TR-BAC in the presence of phorbol myristateacetate (PMA) and hydrocortisone (HC). The inoculum was removed24 h later and the cultures were fed with media containing PMA andHC. At 48 h p.i., the efficiency of infection was determined bystaining coverslips with antibodies to the immediate early proteins(IE1-72 and IE2-86) and the DNA polymerase accessory protein(UL44). In spite of infecting cells at a multiplicity of 5, the percentageof IE-positive (IEþ) cells did not approach 100%; in fact, thepercentage of IEþ cells varied and ranged between 68 and 86%over a series of experiments (Table 1). The difference betweenpercentages of infected cells (IE positivity) for the two strains withina given experiment was generally within 10%. We also determinedthe ratio of UL44þ/IEþ cells to investigate whether there werestrain-specific differences in the transition from IE to E geneexpression in the infected THP-1 derived macrophages. As shownin Table 1, we did not observe any strain-specific differences in thetransition to early gene expression as determined by UL44 expres-sion. In some experiments the UL44þ/IEþ ratio was higher inTowne infected cultures, while it was higher in TR-BAC infectedcultures in other experiments.

We then compared the production of extracellular virus indifferentiated THP1 cells infected with either Towne or TR-BAC.For these experiments, we examined the accumulation of virus incultures from d 5 to d 9 p.i., without replenishing PMA/HC in theculture media. Supernatants were collected until the majority ofcells had lifted from the bottom of the tissue culture vessel. Thekinetics of extracellular virus production are shown in Fig. 1A. Notunexpectedly, the production of virus increased over time andtiter approximately doubled every 24 h; however, we did notobserve differences in extracellular virus production betweenTowne and TR-BAC samples. Consequently, we measured intra-cellular virus production in THP-1 derived macrophages infectedwith either Towne or TR-BAC to test whether there was adifference between the strains in this cell type. Towne and TR-BAC infected cells contained comparable quantities of intracellu-lar virus titers (Fig. 1B). An unexpected observation was thesurvival of the THP-1 derived macrophage-like cells for greaterthan 7 d after infection at a high MOI; however, the extendedperiod of viability and production extracellular virus is similar tothat previously described for primary monocyte-derived

V. Sanchez et al. / Virology 433 (2012) 64–7266

macrophages (Frascaroli et al., 2009; Sinzger et al., 2006) andTHP-1 derived macrophages (McCormick et al., 2010).

Viral DNA accumulates slowly in THP-1 derived macrophages

In addition to analyzing trends in virus production, the kineticsof viral DNA replication between strains was compared. THP-1cells were differentiated as before and infected at high MOI witheither Towne or TR-BAC. At 24 h p.i., the inoculum was removedand cells were collected by trypsinization. Cell pellets werewashed before they were frozen. The quantity of viral DNA atsubsequent time points was measured relative to the level in this24 h sample. At 24 h intervals, cells were harvested and pelletswere stored until the end of the experiment. DNA was extractedfrom cells as previously described (Sanchez and Spector, 2006;White et al., 2004). Viral DNA was measured by real-time PCRusing primers and probe specific for the UL77 gene. DNA sampleswere normalized to the cellular beta-actin gene using commer-cially available primers and probe. Fig. 2A shows the kinetics of

Fig. 2. Viral DNA replication in THP-1 derived macrophages. (A.) THP-1 cells were

treated with PMA to induce differentiation for 22–24 h prior to infection with

HCMV Towne or BAC-derived TR (TR-B) at a multiplicity of 5 as described in the

Materials and Methods. At the time points indicated, infected cells were collected

and frozen. DNA was extracted and used as template for RT-PCR with primers to

HCMV UL77 and human beta-actin as described in the Materials and Methods.

Amplification of viral DNA is shown relative to the quantity present at 24 h p.i.

Results from three independent experiments are shown. (B.) For comparison, the

kinetics of viral DNA replication in permissive human foreskin fibroblasts (HF) is

shown. G0-synchronized fibroblasts were infected with HCMV Towne at an MOI of

5. Replication was measured as described above. (A) Replication in THP- cells, (B)

Replication in human fibroblasts.

viral DNA replication in THP1-infected cells in three independentexperiments. It should be noted that at 24 h p.i., THP-1 macro-phages infected with TR-BAC contained 2–3 fold more genomesthan cells infected with Towne, most likely due to differences inthe particle-to-PFU ratios of the stocks used to infect the cultures.In general, the viral DNA increased less than 10-fold within 72 hp.i. relative to the quantity of viral DNA at 24 h p.i., with theexception of a single experiment with TR-BAC. The slopes of thecurves became steeper after 4 d p.i. The accumulation of viralDNA in THP-1 derived macrophages was slow at early timesrelative to similar experiments in human fibroblasts (Fig. 2B). Infibroblasts infected with the Towne strain, there is an approxi-mately 15-fold increase in viral DNA between 24 and 48 h p.i.Thus we observe a significant delay in viral DNA replication in theTHP-1 derived macrophages, which likely contributes to the delayin extracellular virus production.

The transition to early gene expression is delayed in THP-1 derived

macrophages

We next determined the pattern of viral gene expression byimmunoblotting for proteins of the three kinetic classes. Theresults from representative blots are shown in Fig. 3. The IEproteins were detected within 24 h p.i. We observed that the

Fig. 3. Viral protein expression in THP-1 derived macrophages. THP-1 cells were

treated with PMA to induce differentiation for 22–24 h prior to infection with

HCMV Towne (T) or BAC-derived TR (TR-B) at a multiplicity of 5 as described in

the Materials and Methods. At the time points indicated, infected cells were

collected and frozen. Equivalent cell numbers were loaded into wells of SDS-PAGE

gels. Proteins were resolved and transferred to nitrocellulose as previously

described. Membranes were probed with antibodies to IE (IE1-72 and IE2-86,

-60, and -40), E (UL44, UL57, pp65) and L (pp28) proteins. Antibody to beta-actin

was used as a loading control. Mock-infected cells are indicated in lanes labeled M.

Fig. 4. pp65 localization in THP-1 derived macrophages. THP-1 cells were seeded

V. Sanchez et al. / Virology 433 (2012) 64–72 67

migration patterns of both IE1-72 and IE2-86 differed betweenthe two strains of virus. The antibody CH16.0 recognizes thecommon exon 2 and thus detects both IE proteins (top panel,Fig. 3). We observed a similar shift in size when we probedmembranes with antibodies specific for IE1 or IE2 that recognizethe unique exon 4 or exon 5 of each protein, respectively. The IE2-specific antibody also detects smaller forms of IE2, denoted p40and p60. These smaller forms of IE2 also migrate faster in samplesfrom Towne-infected cells (Fig. 3). The elevated steady state levelsof IE proteins at early times post-infection in cells infected withTR-BAC likely reflect the 2–3 fold greater quantity of genomespresent in the cells. We also noted a change in the IE1/IE2 ratiobetween Towne and TR-BAC infected cells, suggesting differencesin the utilization of IE splice sites within cells infected with eitherstrain (Fig. 3 and Supplementary Table 1).

The expression patterns of the early and early-late proteinsUL44, UL57, and pp65, were also examined. We did not detectsubstantial differences in the expression of these proteinsbetween virus strains at the various time points, except in therates of accumulation of UL44 (Supplementary Table 1). However,by the late time points the levels of UL44 were comparablebetween the strains. Similar observations were made for theexpression of the true late protein pp28. The slow accumulationof early and late proteins is consistent with the delayed onset ofviral DNA replication.

onto glass coverslips and treated with PMA to induce differentiation. Twenty-four

h later, cells were infected with HCMV Towne or BAC-derived TR at a multiplicity

of 5 as described in the Materials and Methods. At 96 and 120 h p.i., cells were

rinsed with PBS and fixed in 2% paraformaldehyde. Coverslips were stained as

previously described with antibodies to HCMV pp65 and the cis-Golgi marker

GM130. Hoechst was used as a nuclear counterstain. Arrows indicate uninfected

cells.

Fig. 5. pp65 localization in primary human fibroblasts. G0-synchronized fibro-

blasts were seeded onto glass coverslips and infected with HCMV Towne or BAC-

derived TR at an MOI of 5, or mock-infected. At 96 h p.i., cells were rinsed with PBS

and fixed in 2% paraformaldehyde. Coverslips were stained as previously

described with antibodies to HCMV pp65 and the cellular protein LAP2. Hoechst

was used as a nuclear counterstain. Confocal images of 0.5 mM sections were

collected along the Z-axis but one single plane through the middle of the cells

is shown.

Cytoplasmic localization of the tegument protein pp65 late in

infection occurs slowly in THP-1 macrophages

As the next step in characterizing the HCMV lytic cycle in THP-1 macrophages, we analyzed the localization of the abundanttegument protein pp65 (UL83). This protein exhibits a biphasicpattern of localization in infected fibroblasts; at early times post-infection, the protein is nuclear while it is exclusively cytoplasmicat late times (Sanchez et al., 2000, 2007). We observed thebiphasic pattern of pp65 localization in infected THP-1 derivedmacrophages, but there was a delay in the transition of pp65 tothe cytoplasm (Fig. 4). In fibroblasts infected with HCMV Towneat a high multiplicity, pp65 is localized to the cytoplasm in themajority of cells at 72 h p.i. (Sanchez et al., 2007; Sanchez et al.,2000). In contrast, in THP-1 macrophages infected with Towne,only 49% contained some pp65 in the cytoplasm at 120 h p.i.(average of 2 independent experiments). At 144 h p.i, pp65 wasdetected in the cytoplasm of approximately 53% in cells infectedwith Towne and approximately 20% of cells no longer containedpp65 in their nuclei. Surprisingly, pp65 appeared to be morestrongly localized to the nuclei of cells infected with TR-BAC incomparison to Towne, as shown in Fig. 4. At 120 h p.i, approxi-mately 23% of TR-BAC-infected THP-1 macrophages containeddetectable levels of pp65 in cytoplasm (average of 2 independentexperiments). The percentage of TR-BAC infected cells containingpp65 in the cytoplasm did increase over time but cells that werecompletely devoid of nuclear pp65 were not detected at 144 h p.i.In order to determine if this was a cell type-specific effect on pp65localization, we examined pp65 localization in Towne and TR-BACinfected primary human fibroblasts. As was previously described,greater than 95% of cells infected with Towne contained pp65exclusively in the cytoplasm at 96 h p.i. (Fig. 5, average of twoexperiments). LAP2 (lamina-associated protein 2) staining wasincluded to demarcate the margins of the nuclei (Dechat et al.,2000). In contrast, 70% of cells infected with TR-BAC containedpp65 in their nuclei at 96 h p.i. (average of 2 experiments),suggesting that the complete evacuation of pp65 from the nucleusthat occurs late in infection may not occur with all HCMV strainsor occurs with different kinetics.

Strain-dependent differences in the morphology of infected THP-1

derived macrophages are correlated with changes in the actin

cytoskeleton

We consistently observed that infection of THP1-derivedmacrophages with TR-BAC but not Towne strain caused the cells

V. Sanchez et al. / Virology 433 (2012) 64–7268

to flatten and elongate starting at 48 h p.i. (Fig. 6). This phenotypewas not evident in cultures infected with Towne until 96 h p.i.,and to a lesser degree. Because our experiments utilized unpur-ified supernatants instead of gradient purified virus, we investi-gated if the changes in cell shape were induced by moleculessecreted into the supernatants, the viral secretome (Streblowet al., 2008). Towne and TR-BAC stocks were subjected to UVirradiation as previously described (Fortunato et al., 2000) andused to infect THP-1 derived macrophages as in Fig. 6. Interest-ingly, we did not observe changes in cell morphology in TR-BACinfected cultures when they were infected with UV-irradiatedvirus (Fig. 7). These results suggest that viral gene expression isrequired for the induction of the elongated morphology adoptedby TR-BAC infected THP1-derived macrophages.

We next investigated the mechanisms that resulted in alteredTHP-1 cell morphology upon infection. We hypothesized thatsuch effects would be mediated by differences in the organizationof the cytoskeleton; therefore, we stained cells infected with

Fig. 6. Morphology of infected THP-1 derived macrophages. THP-1 cells were

treated with PMA to induce differentiation for 22–24 h prior to infection with

HCMV Towne (T) or BAC-derived TR at a multiplicity of 5 as described in the

Materials and Methods. At the time points indicated, phase contrast images of

cultures were recorded.

Fig. 7. UV irradiation of TR-BAC inhibits changes in cell morphology upon

infection. THP-1 cells were treated with PMA to induce differentiation for

22–24 h prior to infection with UV-irradiated HCMV Towne (UV T) or BAC-derived

TR (UV TR-Bac) at a multiplicity of 5 as described in the Materials and Methods.

Non-UV treated virus served as controls. At the time points indicated, phase

contrast images of cultures were recorded.

either strain with FITC-phalloidin, which binds to F-actin andstains actin filaments. Fewer than 5% of mock-infected cellscontained actin filaments (Fig. 8A and Supplementary Figs. 1–3).Instead the phalloidin staining was distributed in a dot-likepattern reminiscent of podosomes (Burger et al., 2011). In con-trast, we readily observed the presence of actin filaments in THP-1 derived macrophages infected with TR-BAC (Fig. 8 and Supple-mentary Figs. 1–3). Cells infected with the Towne strain alsocontained actin filaments but they were thinner and less abun-dant in the cells (Fig. 8 and Supplementary Figs. 1–3). Wequantified the number of THP-1 derived macrophages containingelongated, parallel actin filament bundles and found that thenumber increased over time in cultures infected with either strain(Fig. 8B); however, the percentage of TR-BAC infected THP-1macrophages was greater at all time points analyzed. The greatestdifference occurred at 48 h p.i., when there was an approximately5-fold higher percentage of actin fiber-positive cells in cultures

Fig. 8. Organization of the actin cytoskeleton in HCMV-infected THP-1 derived

macrophages. THP-1 cells were seeded onto glass coverslips and treated with PMA

to induce differentiation. Cells were infected with HCMV Towne or BAC-derived

TR at a multiplicity of 5 as described in the Materials and Methods. At the time

points indicated, cells were rinsed with PBS and fixed in 2% paraformaldehyde.

Coverslips were stained as previously described with antibody to HCMV UL44 and

FITC-labeled phalloidin. Hoechst was used as a nuclear counterstain. Confocal

images of 0.5 mM sections were collected along the Z-axis. (A.) Panels on the left

show images of single planes near the bottoms of the cells where they attach to

the coverslips. Panels on the right show corresponding Z-stack projections

representing merged images of all planes of the field. The white arrowheads

indicate the positions of cells with actin fibers. (B.) The percentage of infected cells

containing actin filaments was quantified at the time points indicated. Squares

and circles indicate the values for Towne- and TR-BAC infected cells, respectively,

for two independent experiments. The line connects the mean values for the

individual time points.

V. Sanchez et al. / Virology 433 (2012) 64–72 69

infected with TR-BAC. Although the percentage increased overtime in Towne samples, the peak level (44% at 120 h p.i.) wasbelow than the lowest percentage observed for TR-BAC infections(58% at 48 h p.i.).

The actin phenotypes observed in the THP-1 derived macro-phages were puzzling since it was previously shown that HCMVinfection of fibroblasts and primary macrophages caused the depo-lymerization of actin stress fibers (Sanchez et al., 2000; Seo et al.,2011; Sharon-Friling et al., 2006). Therefore, we proceeded toanalyze the effects of TR-BAC infection on actin polymerization inprimary fibroblasts. Unlike mock-infected THP-1 macrophages(Fig. 8A), mock-infected fibroblasts contained actin stress fibers(Fig. 9E). Surprisingly, we observed that cells infected with TR-BACcontained thick actin stress fibers (Figs. 9C and D); however, aspreviously reported, primary fibroblasts infected with Towne con-tained few actin fibers (Figs. 9A and B, and Supplementary Fig. 4).These fibers were thinner and less abundant than those observed inTR-BAC infected fibroblasts, a phenotype also observed in THP-1derived macrophages. Again we quantified the number of infectedcells containing elongated, parallel actin bundles and determinedthat actin filaments were depolymerized over time in cells infectedwith Towne (Fig. 9F). In contrast, even at 120 h p.i., greater than 90%of cells infected with TR contained easily discernable actin filaments(Fig. 9F, average of two experiments). Taken together with ourobservations made in THP-1 derived macrophages, these resultssuggest that the degree of actin polymerization during infection isboth cell type-dependent and HCMV strain dependent.

Fig. 9. Organization of the actin cytoskeleton in HCMV-infected human primary foresk

infected with HCMV Towne or BAC-derived TR at a mulitplicity of 5. At the time points

were stained as previously described with antibody to HCMV UL44 and FITC-labeled ph

single planes near the bottoms of the cells where they attach to the coverslips. The white

human fibroblasts; (C, D): TR-BAC infected human fibroblasts; (E): mock infected fibrob

at the time points indicated. Circles indicate the values for Towne- and TR-BAC-infected

time point determined from 2 to 4 independent experiments.

Discussion

Because HCMV remains latent in cells of the myeloid lineage,numerous studies have investigated infections in monocytes andmonocyte-derived macrophages (Chan et al., 2008; Frascaroli et al.,2009; Keyes et al., 2012; Smith et al., 2004; Soderberg-Naucleret al., 1997). Unlike previous studies, we have utilized the THP-1monocytic cell line as a model for lytic replication in macrophages.THP-1 cells have been used primarily as a model to study HCMVreactivation from latency but our studies indicate that they mayalso be useful for examining aspects of productive viral infection.One advantage of using these cells is that large numbers of cells canbe propagated easily as opposed to isolating primary monocytesfrom peripheral blood. Secondly, experiments can be conductedwithin a single genetic background, which may not necessarily bethe case when isolating monocytes from donor blood products. Themajor caveat to using these cells is their lack of functional p53(Auwerx, 1991), which has been shown to be required for efficientinfection in fibroblasts (Casavant et al., 2006); however, ourcomparative studies presented here indicate that infections inTHP1-derived macrophages resemble infections in primary fibro-blasts, although we did observe some cell type-specific differencesin actin organization. In addition to the results shown in this report,we have previously demonstrated that the ABCA1 transporter isdown-regulated primary fibroblasts and THP-1 derived macro-phages (Sanchez and Dong, 2010), further validating the use ofTHP-1 macrophages as a model for infection of primary cells.

in fibroblasts. G0-synchronized fibroblasts were seeded onto glass coverslips and

indicated, cells were rinsed with PBS and fixed in 2% paraformaldehyde. Coverslips

alloidin. Hoechst was used as a nuclear counterstain. (A–E) Panels show images of

arrowheads indicate the positions of cells with actin fibers. (A, B): Towne-infected

lasts. (F.) The percentage of infected cells containing actin filaments was quantified

cells for each independent experiment. Columns indicate the mean value for each

V. Sanchez et al. / Virology 433 (2012) 64–7270

While cell cultures are treated with PMA to induce differentia-tion one day prior to infection, we observed that a subset of cellsdid not express IE proteins 48 h p.i., despite infection at highmultiplicity. The percentage of cells that did not express IEproteins varied between 15 and 35% over multiple experiments(Table 1), which may reflect differences in the rate of develop-ment of the THP-1 monocytes into macrophage-like cells. McCor-mick and colleagues reported that surface expression of CD11band CD14, markers of macrophage differentiation, increased byday 3 post differentiation of the THP-1 cells (McCormick et al.,2010). It is possible that altering the protocol for inducingdifferentiation may increase the efficiency of infection; in fact, itwas recently shown that THP-1 cultures that were allowed to restfor 5 days after treatment with phorbol esters to induce differ-entiation displayed multiple markers and activities that madethem indistinguishable from primary macrophages (Daigneaultet al., 2010).

As shown in Fig. 2, there was an eclipse period followinginfection before there was a significant increase in viral DNAreplication in THP-1 derived macrophages. It is unclear if thedelay in replication also occurs in primary monocyte-derivedmacrophages or results from the absence of functional p53 inTHP-1 cells. Casavant and colleagues reported slow accumulationof viral DNA in p53-negative fibroblasts relative to wild type cells(Casavant et al., 2006). The steady state levels of the earlyproteins UL44 and UL57 that are necessary for replicationremained low until 72 h p.i. in THP-1 macrophages, consistentwith the delay in replication. By 120 h p.i., the viral protein levelswere similar between samples infected with each strain. At thistime point, comparable amounts of virus were also released bycells infected with Towne or TR-BAC. Because TR-BAC is a clinicalstrain, we expected that it would replicate more efficiently thanTowne in the THP-1 derived macrophages but we did not observea marked difference between the strains. As mentioned above,viral replication is slow in p53-negative fibroblasts (Casavantet al., 2006); therefore, the p53 mutation in THP1 cells may maskany growth advantage that TR-BAC may have. Alternatively, theresult that the two strains are similar in terms of extracellularvirus production in the THP-1 macrophage model, but not infibroblasts, is reminiscent of previous observations by Adler andcolleagues. It has been shown that the presence of wild typeUL131A in clinical strains has an inhibitory effect on virus releasein fibroblasts but not in endothelial cells (Adler et al., 2006). Wenoted that the TR-BAC infected fibroblasts do not produce asmuch extracellular virus as Towne-infected fibroblasts (data notshown), confirming the defect in virus release from these cells.Thus our results in THP1 derived macrophages may reflect theeffects of UL131A on TR-BAC release, as well as the p53 defect inthis cell type.

While there were some modest differences in the accumula-tion of IE proteins in cells infected with Towne and TR-BAC atearly times (Fig. 3), the most striking observation was that the IEproteins (full length IE1, IE2, and IE2 p60 and IE2 p40) migratedmore slowly in TR-BAC samples. Analysis of UL122–123 sequencein TR-BAC showed a small number of sequence changes incomparison to the Towne strain, which could alter the migrationof the proteins. It is also possible that migration was affected bydifferences in the pattern of post-translational modifications thatoccur in exons 2, 3, 4, and 5. We also observed that the IE1/2 ratiowas different between the two strains at late times (Supplemen-tary Table 1), but software analysis predicted that similar splicesites were present in the sequences. Thus, other factors likelyinfluence splice site-utilization in the UL122-123 locus.

Although we did not observe major differences in the growthcharacteristics of the laboratory strain Towne versus the clinicalstrain TR-BAC in THP-1 derived macrophages, some distinct

morphological changes were noted in cells infected with TR-BAC. These results were correlated with the presence of actinfilaments in TR-BAC-infected THP-1 macrophages at higher fre-quency than in Towne-infected cells. It has been well documentedthat infection with AD169, Towne, and Toledo strains results indisruption of the actin cytoskeleton and/or focal adhesions infibroblasts raising the possibility that we were observing eithercell type-specific and/or virus-strain specific differences in actinmobilization in the THP-1 derived macrophages (Sanchez et al.,2000; Seo et al., 2011; Sharon-Friling et al., 2006; Stanton et al.,2007). Additionally, Frascaroli and coworkers reported that theactin cytoskeleton was reorganized in primary macrophagesinfected with HCMV TB40E but they did not report the presenceof actin fibers in infected cells (Frascaroli et al., 2009). Theirobservations raised the possibility that our results in THP-1macrophages were an artifact of using a transformed myeloidcell type. However, fibroblasts infected with TR-BAC also con-tained abundant actin filaments while those infected with Townestrain did not late in infection, indicating that virus strains differin their ability to remodel the cytoskeleton even in primary cells.We postulate that the difference in the kinetics of actin poly-merization between Towne-infected THP-1 macrophages andTowne-infected fibroblasts may reflect the difference in the initialstate of the actin cytoskeleton in each cell type before infection.The virus strains may activate distinct signaling pathways thatinfluence the ability of infected cells to remodel the actincytoskeleton in a cell type-dependent manner. Additional experi-ments are necessary to examine these possibilities.

Interestingly, we also observed that pp65 was retained in thenuclei of TR-BAC infected THP-1 derived macrophages. In addi-tion, the localization of pp65 to the cytoplasm was slow inTowne-infected THP-1 macrophages relative to primary fibro-blasts. Casavant et al. reported a delay in the transition of pp65 tothe cytoplasm in Towne-infected p53-null fibroblasts (Casavantet al., 2006), so it is possible that the delay in THP-1 macrophagescan be explained the absence of functional p53 in these cells.Nevertheless, the strain-specific differences in pp65 localizationwe observed in THP-1 derived macrophages were also detected inprimary fibroblasts. Nuclear export of pp65 from the nucleus isregulated by cyclin-dependent kinase (cdk) activity (Sanchezet al., 2007; Sanchez and Spector, 2006), which may suggest thatthe TR-BAC strain of virus does not activate cdks in the samemanner as Towne.

Taken together, our work in THP-1 macrophages and fibro-blasts has revealed novel strain-specific differences in the abilityof HCMV to remodel the cytoskeleton and to regulate tegumentprotein localization. Thus, these studies validate the use ofdifferentiated THP-1 cells as a model for lytic infection.

Materials and methods

Cell culture and virus

HCMV virus strains were propagated in primary humanfibroblasts as previously described (Tamashiro et al., 1982). Thelaboratory strain Towne was obtained from American TypeCulture Collection (ATCC; VR 977). The TR-BAC was preparedfrom the TR clinical HCMV isolate (a gift from Dr. Jay Nelson,Oregon Health Sciences University). The BAC clone was generatedby substituting BAC DNA for the US2–US5 region of the TRgenome (Murphy et al., 2003). It should be noted that stocks ofvirus were prepared in fibroblasts not endothelial cells as pre-viously described (Sinzger et al., 2006).

THP-1 cells (ATCC) were maintained in RPMI Medium 1640supplemented with 10% heat-inactivated fetal bovine serum,

V. Sanchez et al. / Virology 433 (2012) 64–72 71

glutamine, penicillin and streptomycin. Cells were counted,seeded into tissue culture plates, and differentiated 22–24 hbefore infection by treatment with phorbol myristate acetate(PMA; EMD Biosciences) at a final concentration of 50 ng/mL.Adherent cells were infected with HCMV Towne or TR-BAC at amultiplicity of infection (MOI) of 5 in the presence of PMA andhydrocortisone (HC; 5 mM; Sigma). Twenty-four h p.i., the virusinoculum was removed and cultures were fed with completemedium containing PMA and HC. Cells were harvested at varioustimes p.i. and pellets were frozen until further analysis. Alterna-tively, supernatants containing progeny virions were collected atthe times indicated and viral titers were determined on primaryhuman fibroblasts as previously described (Tamashiro et al.,1982). For some experiments, cell pellets were collected alongwith supernatants for measurement of intracellular virus titer.Briefly, pelleted cells were resuspended in complete media anddrawn through a sterile syringe with a 25-gauge needle. Thelysate was gently sonicated in a water bath sonicator for 2 minthen frozen. Viral titer was determined on primary humanfibroblasts as described above.

Human foreskin fibroblasts (HFF) were maintained as pre-viously described in minimal essential media (MEM)-Earle’smedia supplemented with 10% fetal bovine serum, penicillin,streptomycin, amphotericin, and glutamine (Sanchez et al.,2002). For G0 infections, cells were grown to confluence allowedto arrest for 3 additional days before trypsinization and replatingat a lower density (Sanchez et al., 2003). Cells grown on coverslipswere infected at a MOI of 5 with Towne or TR-BAC the time ofrelease from confluence. Mock-infected cells were incubated withan equal volume of conditioned media.

UV inactivation of virus

To determine the effects of the viral secretome on THP-1 cellmorphology, the virus inoculum was inactivated by UV irradia-tion as previously described (Fortunato et al., 2000). An equiva-lent number of plaque forming units (PFU) Towne or TR-BACstocks were diluted in RPMI and exposed to 6000 J/m2 of irradia-tion in a Stratalinker. Sodium pyruvate (5 mM), PMA, and HCwere added to the inoculum after UV treatment, which was thenused to infect THP-1 derived macrophages 24 h post-differentia-tion as described above. Cells were infected with Towne orTR-BAC in parallel. At 48, 72, 96, and 120 h p.i., cells werevisualized by phase contrast microscopy and images wererecorded.

Western blotting

At the time points indicated, HCMV- and mock-infected cellswere collected by trypsinization and frozen. Pellets were resus-pended in reducing sample buffer (RSB) (Sanchez et al., 2002) anddisrupted by passing samples several times through a 25-gaugeneedle. Samples were boiled for 3 min prior to loading. Equivalentcell numbers were loaded into each lane of polyacrylamide gelsand proteins were resolved by SDS-PAGE. Proteins were trans-ferred to nitrocellulose. The following antibodies were used toprobe filters: anti-IE1/IE2, anti-UL44, and anti-UL57 (Virusys);anti-IE1, anti-pp65, anti-pp28 (a gift from Dr. William Britt,University of Alabama, Birmingham); anti-IE2 (Chemicon); andanti-beta actin (Sigma).

Analysis of viral DNA replication by real time PCR

THP-1 derived macrophages were infected with Towne orTR-BAC at an MOI of 5 as described above. Cells were collectedat 24, 48, 72, 96, 120, and 144 h p.i. and frozen. To measure viral

DNA replication in fibroblasts, cells were infected as describedabove and infected with Towne at a MOI of 5. Cells were collectedand pellets were frozen at the time points indicated. DNA wasextracted from the pellets using the Qiagen Mini-Blood Kit permanufacturer’s instructions. Viral DNA replication was measuredby quantitative real time PCR using TaqMan One-Step PCR mastermix reagents kit (Applied Biosystems) and primers to HCMV UL77(White et al., 2004), TaqMan dual-labeled (50 fluorescein (FAM) -30 Black Hole quencher) probes (Integrated DNA Technologies) asdescribed previously (White et al., 2004). A standard curve wasmade using dilutions of the DNA from samples harvested at 24 hp.i. Fold differences were calculated relative to the level of viralDNA at 24 h p.i. Samples were normalized by measuring theamplification of the cellular beta actin gene as an internalstandard. Primers and probe for beta actin were purchased fromApplied Biosystems.

Immunofluorescence

THP-1 monocytes were seeded onto glass coverslips andtreated with PMA to induced differentiation 22–24 h prior toinfection Towne or TR-BAC at an MOI of 5 as described above. Forinfections in primary human fibroblasts, G0-synchronized cellswere seeded onto glass coverslips and infected at a MOI of 5.At the time points indicated, cells were rinsed with PBS and fixedin 2% paraformaldehyde in PBS for 15 min at room temperature.Cells were permeabilized and stained as previously described(Sanchez et al., 2002). Coverslips were stained with anti-pp65antibody (a gift from Dr. William Britt); anti-UL44 antibody(Virusys); anti-GM130 and anti-LAP2 antibodies (BD Biosciences);and FITC-phalloidin (Sigma). Coverslips were mounted withSlowFade reagent (Invitrogen) and visualized using an OlympusIX81 research microscope equipped with a DSU spinning diskconfocal unit. Images were captured by either epifluorescent orconfocal imaging techniques as indicated in the figure legends.For measurement of the number of cells containing pp65 in nucleiversus cytoplasm, 5–10 images were recorded per coverslip andapproximately 100 cells were counted per experimental condi-tion. For quantitation of cells containing actin fibers, 8–18 imageswere recorded per coverslip (approximately 100 cells on averageper experimental condition). Data from 2–4 independent experi-ments were analyzed.

Acknowledgments

The authors would like to thank Dr. Jay Nelson for providingTR-BAC virus and Dr. William Britt for antibodies and valuablediscussions. We also thank Dr. Vernon Tesh at Texas A&M HealthScience Center for providing THP-1 monocytes. This work wassupported by the Texas A&M Health Science Center College ofMedicine.

Appendix A. Supporting information

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.virol.2012.07.015.

References

Adler, B., Scrivano, L., Ruzcics, Z., Rupp, B., Singzer, C., Koszinowski, U., 2006. Roleof human cytomegalovirus UL131A in cell-type virus entry and release. J. Gen.Virol. 87 (9), 2451–2460.

Akashi, M., Osawa, Y., Koeffler, H.P., Hachiya, M., 1999. p21Waf1 expression by anactivator of protein kinase C is regulated mainly at the post-transcriptional

V. Sanchez et al. / Virology 433 (2012) 64–7272

level in cells lacking p53: important role of RNA stabilization. Biochem. J. 337,606–616.

Auwerx, J., 1991. The human leukemia cell line, THP-1: a multifaceted model forthe study of monocyte-macrophage differentiation. Experentia 47, 22–31.

Burger, K.L., Davis, A.L., Isom, S., Mishra, N., Seals, D.F., 2011. The podosome markerprotein Tks5 regulates macrophage invasive behavior. Cytoskeleton 68,694–711.

Casavant, N.C., Luo, M.H., Rosenke, K., Winegardner, T., Zurawska, A., Fortunato,E.A., 2006. Potential role for p53 in the permissive life cycle of humancytomegalovirus. J. Virol. 80 (17), 8390–8401.

Chan, G., Bivins-Smith, E.R., Smith, M.S., Smith, P.M., Yurochko, A.D., 2008.Transcriptome analysis reveals human cytomegalovirus reprograms monocytedifferentiation toward an M1 macrophage. J. Immunol. 181, 698–711.

Daigneault, M., Preston, J.A., Marriott, H.M., Whyte, M.K.B., Dockrell, D.H., 2010.The identification of markers of macrophage differentiation in PMA-stimu-lated THP-1 cells and monocyte-derived macrophages. PLoS ONE 5 (1), e8668, ,http://dx.doi.org/10.1371/journal.pone.0008668.

Dechat, T., Korbei, B., Vaughan, O.A., Vicek, S., Hutchison, C.J., Foisner, R., 2000.Lamina-associate polypeptide 2alpha binds intranuclear A-type lamins. J. CellSci. 113 (19), 3473–3484.

Fortunato, E.A., Dell’Aquila, M.L., Spector, D.H., 2000. Specific chromosome 1breaks induced by human cytomegalovirus. Proc. Natl. Acad. Sci. USA 97 (2),853–858.

Frascaroli, G., Varani, S., Blankenhorn, N., Pretsch, R., Bacher, M., Leng, L., Bucala, R.,Landini, M.P., Mertens, T., 2009. Human cytomegalovirus paralyzes macro-phage motility through down-regulation of chemokine receptors, reorganiza-tion of the cytoskeleton, and release of macrophage migration inhibitoryfactor. J. Immunol. 182, 477–488.

Ioudinkova, E., Arcangeletti, M.C., Rynditch, A., Conto, F.D., Motta, F., Covan, S.,Pinardi, F., Razin, S.V., Chezzi, C., 2006. Control of human cytomegalovirusgene expression by differential histone modification during lytic and latentinfection of a monocytic cell line. Gene 384, 120–128.

Keyes, L.R., Bego, M.G., Soland, M., Jeor St., S., 2012. Cyclophilin A (CyPA) isrequired for efficient HCMV DNA replication and reactivation. J. Gen. Virol. 93(4), 722–732.

Kohro, T., Tanaka, T., Murakami, T., Wada, Y., Aburatani, H., Hamakubo, T., Kodama,T., 2004. A comparison of differences in the gene expression profiles of phorbol12-myristate differentiated THP-1 cells and human monocyte-derived macro-phages. J. Atheroscler. Thromb. 11, 88–97.

Martinez, F.O., Gordon, S., Locati, M., Mantovani, A., 2006. Transcriptional profilingof the human monocyte-to-macrophage differentiation and polarization: newmolecules and patterns of gene expression. J. Immunol. 177, 7303–7311.

McCormick, A.L., Roback, L., Livingston-Rosanoff, D., Clair St., C., 2010. The humancytomegalovirus UL36 gene controls caspase-dependent and -independent celldeath program activated by infection of monocytes differentiating to macro-phages. J. Virol 84 (10), 5108–5123.

Murphy, E., Yu, D., Grimwood, J., Schmutz, J., Dickson, M., Jarvis, M.A., Nelson, J.A.,Myers, R.M., Shenk, T.E., 2003. Coding potential of laboratory and clinicalstrains of human cytomegalovirus. Proc. Natl. Acad. Sci. USA 100 (25),14976–14981.

Murray, P.J., Wynn, T.A., 2011. Protective and pathogenic functions of macrophagesubsets. Nat. Rev. Immunol. 11, 723–737.

Reeves, M., Sinclair, J., 2008. Aspects of human cytomegalovirus latency andreactivation. Curr. Top. Microbiol. Immunol. 325, 297–3131.

Saffert, R.T., Kalejta, R.F., 2010. Cellular and viral control over the initial events ofhuman cytomegalovirus experimental latency in CD34þ cells. J. Virol. 84 (11),5594–5604.

Sanchez, V., Clark, C.L., Yen, J.Y., Dwarakanath, R., Spector, D.H., 2002. Viablehuman cytomegalovirus recombinant virus with an internal deletion of the IE286 gene affects late stages of viral replication. J. Virol. 76 (6), 2973–2989.

Sanchez, V., Dong, J.J., 2010. Alteration of lipid metabolism in cells infected withhuman cytomegalovirus. Virology 404 (1), 71–77.

Sanchez, V., Mahr, J.A., Orazio, N.I., Spector, D.H., 2007. Nuclear export of thehuman cytomegalovirus tegument protein pp65 requires cyclin-dependentkinase activity and the Crm1 exporter. J. Virol. 81 (21), 11730–11736.

Sanchez, V., McElroy, A.K., Spector, D.H., 2003. Mechanisms governing mainte-nance of cdk1/cyclin B1 kinase activity in cells infected with human cytome-galovirus. J. Virol. 77 (24), 13124–13224.

Sanchez, V., Spector, D.H., 2006. Cyclin-dependent kinase activity is required forefficient expression and post-translational modification of human cytomegalovirusproteins and for production of extracellular particles. J. Virol. 80 (12), 5886–5896.

Sanchez, V., Greis, K.D., Sztul, E., Britt, W.J., 2000. Accumulation of virion tegumentand envelope proteins in a stable cytoplasmic compartment during humancytomegalovirus replication: characterization of a potential site of virusassembly. J. Virol. 74 (2), 975–986.

Seo, J.-Y., Yaneva, R., Hinson, E.R., Creswell, P., 2011. Human cytomegalovirusdirectly induces the antiviral protein viperin to enhance infectivity. Science332, 1093–1097.

Sharon-Friling, R., Goodhouse, J., Colberg-Poley, A.M., Shenk, T., 2006. Humancytomegalovirus pUL37�1 induces the release of endoplasmic reticulumcalcium stores. Proc. Natl. Acad. Sci. USA 103 (50), 19117–19122.

Sinclair, J.H., Baillie, J., Bryant, L.A., Taylor-Wiedeman, J.A., Sissons, J.G., 1992.Repression of human cytomegalovirus major immediate early gene expressionin a monocytic cell line. J. Gen. Virol. 73 (2), 433–435.

Sinzger, C., Eberhardt, K., Cavignac, Y., Weinstock, C., Kessler, T., Jahn, G., Davignon,J.L., 2006. Macrophage cultures are susceptible to lytic productive infection byendothelial-cell-propagated human cytomegalovirus strains and present viralIE1 protein to CD4þ T cells despite late downregulation of MHC class IImolecules. J. Gen. Virol. 87, 1853–1862.

Smith, M.S., Bentz, G.L., Alexander, J.S., Yurochko, A.D., 2004. Human cytomega-lovirus induces monocyte differentiation and migration as a strategy fordissemination and persistence. J. Virol. 78 (9), 4444–4453.

Soderberg-Naucler, C., Fish, K.N., Nelson, J.A., 1997. Reactivation of latent humancytomegalovirus by allogeneic stimulation of blood cells from healthy donors.Cell 91 (1), 119–126.

Stanton, R.J., McSharry, B.P., Rickards, C.R., Wang, E.C.Y., Tomasec, P., Wilkinson,G.W., 2007. Cytomegalovirus destruction of focal adhesions revealed in a high-throughput western blot analysis of cellular protein expression. J. Virol. 81(15), 7860–7872.

Streblow, D.N., Dumortier, J., Moses, A.V., Orloff, S.L., Nelson, J.A., 2008. Mechan-isms of cytomegalovirus-accelerated vascular disease: induction of paracrinefactors that promote angiogenesis and wound healing. Curr. Top. Microbiol.Immunol. 325, 397–415.

Tamashiro, J.C., Hock, L.J., Spector, D.H., 1982. Construction of a cloned library ofthe EcoRI fragments from the human cytomegalovirus genome (strain AD169).J. Virol. 42, 547–557.

Turtinen, L.W., Seufzer, B.J., 1994. Selective permissiveness of TPA differentiatedTHP-1 myelomonocytic cell cultures. Microb. Pathog. 16 (5), 373–378.

White, E.A., Clark, C.L., Sanchez, V., Spector, D.H., 2004. Small internal deletion inthe human cytomegalovirus IE2 gene result in nonviable recombinant viruseswith differential defects in viral gene expression. J. Virol. 78 (4), 1817–1830.

Yee, L.-F., Lin, P.L., Stinski, M.F., 2007. Ectopic expression of HCMV IE71 and IE86proteins is sufficient to induce early gene expression but not production ofinfectious virus in undifferentiated promonocytic THP-1 cells. Virology 363,174–188.